UK Road Safety: Traffic Accidents (2005-2017)

1. Explanation

1.1. Brief

Welcome!

For this EDA, this research will use UK Road Safety: Traffic Accidents (2005-2017).

reference: https://www.kaggle.com/tsiaras/uk-road-safety-accidents-and-vehicles

1.2. Data’s Point of View

These files provide detailed road safety data about the circumstances of personal injury road accidents in GB from 1979, the types of vehicles involved and the consequential casualties. The statistics relate only to personal injury accidents on public roads that are reported to the police, and subsequently recorded, using the STATS19 accident reporting form.

2. Data Inspection

Library

library("lubridate")
library("dplyr")

Data source:

traffic<-read.csv("Accident_Information.csv")

First 6 data

head(traffic)

last 6 data

tail(traffic)

Data dimension

dim(traffic)

## [1] 2047256      34

Column Names

names(traffic)

##  [1] "Accident_Index"                             
##  [2] "X1st_Road_Class"                            
##  [3] "X1st_Road_Number"                           
##  [4] "X2nd_Road_Class"                            
##  [5] "X2nd_Road_Number"                           
##  [6] "Accident_Severity"                          
##  [7] "Carriageway_Hazards"                        
##  [8] "Date"                                       
##  [9] "Day_of_Week"                                
## [10] "Did_Police_Officer_Attend_Scene_of_Accident"
## [11] "Junction_Control"                           
## [12] "Junction_Detail"                            
## [13] "Latitude"                                   
## [14] "Light_Conditions"                           
## [15] "Local_Authority_.District."                 
## [16] "Local_Authority_.Highway."                  
## [17] "Location_Easting_OSGR"                      
## [18] "Location_Northing_OSGR"                     
## [19] "Longitude"                                  
## [20] "LSOA_of_Accident_Location"                  
## [21] "Number_of_Casualties"                       
## [22] "Number_of_Vehicles"                         
## [23] "Pedestrian_Crossing.Human_Control"          
## [24] "Pedestrian_Crossing.Physical_Facilities"    
## [25] "Police_Force"                               
## [26] "Road_Surface_Conditions"                    
## [27] "Road_Type"                                  
## [28] "Special_Conditions_at_Site"                 
## [29] "Speed_limit"                                
## [30] "Time"                                       
## [31] "Urban_or_Rural_Area"                        
## [32] "Weather_Conditions"                         
## [33] "Year"                                       
## [34] "InScotland"

3. Data Cleansing & Coercion

check variable type

str(traffic)

## 'data.frame':    2047256 obs. of  34 variables:
##  $ Accident_Index                             : chr  "200501BS00001" "200501BS00002" "200501BS00003" "200501BS00004" ...
##  $ X1st_Road_Class                            : chr  "A" "B" "C" "A" ...
##  $ X1st_Road_Number                           : int  3218 450 0 3220 0 0 0 315 3212 450 ...
##  $ X2nd_Road_Class                            : chr  NA "C" NA NA ...
##  $ X2nd_Road_Number                           : int  0 0 0 0 0 0 0 0 304 0 ...
##  $ Accident_Severity                          : chr  "Serious" "Slight" "Slight" "Slight" ...
##  $ Carriageway_Hazards                        : chr  "None" "None" "None" "None" ...
##  $ Date                                       : chr  "2005-01-04" "2005-01-05" "2005-01-06" "2005-01-07" ...
##  $ Day_of_Week                                : chr  "Tuesday" "Wednesday" "Thursday" "Friday" ...
##  $ Did_Police_Officer_Attend_Scene_of_Accident: int  1 1 1 1 1 1 1 1 1 1 ...
##  $ Junction_Control                           : chr  "Data missing or out of range" "Auto traffic signal" "Data missing or out of range" "Data missing or out of range" ...
##  $ Junction_Detail                            : chr  "Not at junction or within 20 metres" "Crossroads" "Not at junction or within 20 metres" "Not at junction or within 20 metres" ...
##  $ Latitude                                   : num  51.5 51.5 51.5 51.5 51.5 ...
##  $ Light_Conditions                           : chr  "Daylight" "Darkness - lights lit" "Darkness - lights lit" "Daylight" ...
##  $ Local_Authority_.District.                 : chr  "Kensington and Chelsea" "Kensington and Chelsea" "Kensington and Chelsea" "Kensington and Chelsea" ...
##  $ Local_Authority_.Highway.                  : chr  "Kensington and Chelsea" "Kensington and Chelsea" "Kensington and Chelsea" "Kensington and Chelsea" ...
##  $ Location_Easting_OSGR                      : int  525680 524170 524520 526900 528060 524770 524220 525890 527350 524550 ...
##  $ Location_Northing_OSGR                     : int  178240 181650 182240 177530 179040 181160 180830 179710 177650 180810 ...
##  $ Longitude                                  : num  -0.191 -0.212 -0.206 -0.174 -0.157 ...
##  $ LSOA_of_Accident_Location                  : chr  "E01002849" "E01002909" "E01002857" "E01002840" ...
##  $ Number_of_Casualties                       : int  1 1 1 1 1 1 1 2 2 5 ...
##  $ Number_of_Vehicles                         : int  1 1 2 1 1 2 2 1 2 2 ...
##  $ Pedestrian_Crossing.Human_Control          : int  0 0 0 0 0 0 0 0 0 0 ...
##  $ Pedestrian_Crossing.Physical_Facilities    : int  1 5 0 0 0 0 0 0 5 8 ...
##  $ Police_Force                               : chr  "Metropolitan Police" "Metropolitan Police" "Metropolitan Police" "Metropolitan Police" ...
##  $ Road_Surface_Conditions                    : chr  "Wet or damp" "Dry" "Dry" "Dry" ...
##  $ Road_Type                                  : chr  "Single carriageway" "Dual carriageway" "Single carriageway" "Single carriageway" ...
##  $ Special_Conditions_at_Site                 : chr  "None" "None" "None" "None" ...
##  $ Speed_limit                                : int  30 30 30 30 30 30 30 30 30 30 ...
##  $ Time                                       : chr  "17:42" "17:36" "00:15" "10:35" ...
##  $ Urban_or_Rural_Area                        : chr  "Urban" "Urban" "Urban" "Urban" ...
##  $ Weather_Conditions                         : chr  "Raining no high winds" "Fine no high winds" "Fine no high winds" "Fine no high winds" ...
##  $ Year                                       : int  2005 2005 2005 2005 2005 2005 2005 2005 2005 2005 ...
##  $ InScotland                                 : chr  "No" "No" "No" "No" ...

3.1 Data Coertion

Change to Factor/ Categorical

traffic$X1st_Road_Class <- as.factor(traffic$X1st_Road_Class)
traffic$X1st_Road_Number <- as.factor(traffic$X1st_Road_Number)
traffic$X2nd_Road_Class <- as.factor(traffic$X2nd_Road_Class)
traffic$X2nd_Road_Number <- as.factor(traffic$X2nd_Road_Number)
traffic$Accident_Severity <- as.factor(traffic$Accident_Severity)
traffic$Carriageway_Hazards <- as.factor(traffic$Carriageway_Hazards)
traffic$Day_of_Week <- as.factor(traffic$Day_of_Week)
traffic$Did_Police_Officer_Attend_Scene_of_Accident <- as.factor(traffic$Did_Police_Officer_Attend_Scene_of_Accident)
traffic$Junction_Control <- as.factor(traffic$Junction_Control)
traffic$Junction_Detail <- as.factor(traffic$Junction_Detail)
traffic$Light_Conditions <- as.factor(traffic$Light_Conditions)
traffic$Local_Authority_.District. <- as.factor(traffic$Local_Authority_.District.)
traffic$Local_Authority_.Highway. <- as.factor(traffic$Local_Authority_.Highway.)
traffic$Location_Easting_OSGR <- as.factor(traffic$Location_Easting_OSGR)
traffic$Location_Northing_OSGR <- as.factor(traffic$Location_Northing_OSGR)
traffic$LSOA_of_Accident_Location <- as.factor(traffic$LSOA_of_Accident_Location)
traffic$Police_Force <- as.factor(traffic$Police_Force)
traffic$Road_Surface_Conditions <- as.factor(traffic$Road_Surface_Conditions)
traffic$Road_Type <- as.factor(traffic$Road_Type)
traffic$Special_Conditions_at_Site <- as.factor(traffic$Special_Conditions_at_Site)
traffic$Speed_limit <- as.factor(traffic$Speed_limit)
traffic$Urban_or_Rural_Area <- as.factor(traffic$Urban_or_Rural_Area)
traffic$Weather_Conditions <- as.factor(traffic$Weather_Conditions)
traffic$InScotland <- as.factor(traffic$InScotland)

Change to Date

traffic$Date <- ymd(traffic$Date)
traffic$Time <- hm(traffic$Time)

Check variable type again

glimpse(traffic)

## Rows: 2,047,256
## Columns: 34
## $ Accident_Index                              <chr> "200501BS00001", "20050...
## $ X1st_Road_Class                             <fct> A, B, C, A, Unclassifie...
## $ X1st_Road_Number                            <fct> 3218, 450, 0, 3220, 0, ...
## $ X2nd_Road_Class                             <fct> NA, C, NA, NA, NA, NA, ...
## $ X2nd_Road_Number                            <fct> 0, 0, 0, 0, 0, 0, 0, 0,...
## $ Accident_Severity                           <fct> Serious, Slight, Slight...
## $ Carriageway_Hazards                         <fct> None, None, None, None,...
## $ Date                                        <date> 2005-01-04, 2005-01-05...
## $ Day_of_Week                                 <fct> Tuesday, Wednesday, Thu...
## $ Did_Police_Officer_Attend_Scene_of_Accident <fct> 1, 1, 1, 1, 1, 1, 1, 1,...
## $ Junction_Control                            <fct> Data missing or out of ...
## $ Junction_Detail                             <fct> Not at junction or with...
## $ Latitude                                    <dbl> 51.48910, 51.52007, 51....
## $ Light_Conditions                            <fct> Daylight, Darkness - li...
## $ Local_Authority_.District.                  <fct> Kensington and Chelsea,...
## $ Local_Authority_.Highway.                   <fct> Kensington and Chelsea,...
## $ Location_Easting_OSGR                       <fct> 525680, 524170, 524520,...
## $ Location_Northing_OSGR                      <fct> 178240, 181650, 182240,...
## $ Longitude                                   <dbl> -0.191170, -0.211708, -...
## $ LSOA_of_Accident_Location                   <fct> E01002849, E01002909, E...
## $ Number_of_Casualties                        <int> 1, 1, 1, 1, 1, 1, 1, 2,...
## $ Number_of_Vehicles                          <int> 1, 1, 2, 1, 1, 2, 2, 1,...
## $ Pedestrian_Crossing.Human_Control           <int> 0, 0, 0, 0, 0, 0, 0, 0,...
## $ Pedestrian_Crossing.Physical_Facilities     <int> 1, 5, 0, 0, 0, 0, 0, 0,...
## $ Police_Force                                <fct> Metropolitan Police, Me...
## $ Road_Surface_Conditions                     <fct> Wet or damp, Dry, Dry, ...
## $ Road_Type                                   <fct> Single carriageway, Dua...
## $ Special_Conditions_at_Site                  <fct> None, None, None, None,...
## $ Speed_limit                                 <fct> 30, 30, 30, 30, 30, 30,...
## $ Time                                        <Period> 17H 42M 0S, 17H 36M ...
## $ Urban_or_Rural_Area                         <fct> Urban, Urban, Urban, Ur...
## $ Weather_Conditions                          <fct> Raining no high winds, ...
## $ Year                                        <int> 2005, 2005, 2005, 2005,...
## $ InScotland                                  <fct> No, No, No, No, No, No,...

traffic

3.2 Missing Value

Showed any column that has NULL values

colSums(is.na(traffic))

##                              Accident_Index 
##                                           0 
##                             X1st_Road_Class 
##                                           0 
##                            X1st_Road_Number 
##                                           2 
##                             X2nd_Road_Class 
##                                      844272 
##                            X2nd_Road_Number 
##                                       17593 
##                           Accident_Severity 
##                                           0 
##                         Carriageway_Hazards 
##                                           0 
##                                        Date 
##                                           0 
##                                 Day_of_Week 
##                                           0 
## Did_Police_Officer_Attend_Scene_of_Accident 
##                                         278 
##                            Junction_Control 
##                                           0 
##                             Junction_Detail 
##                                           0 
##                                    Latitude 
##                                         174 
##                            Light_Conditions 
##                                           0 
##                  Local_Authority_.District. 
##                                           0 
##                   Local_Authority_.Highway. 
##                                           0 
##                       Location_Easting_OSGR 
##                                         164 
##                      Location_Northing_OSGR 
##                                         164 
##                                   Longitude 
##                                         175 
##                   LSOA_of_Accident_Location 
##                                           0 
##                        Number_of_Casualties 
##                                           0 
##                          Number_of_Vehicles 
##                                           0 
##           Pedestrian_Crossing.Human_Control 
##                                        2920 
##     Pedestrian_Crossing.Physical_Facilities 
##                                        3560 
##                                Police_Force 
##                                           0 
##                     Road_Surface_Conditions 
##                                           0 
##                                   Road_Type 
##                                           0 
##                  Special_Conditions_at_Site 
##                                           0 
##                                 Speed_limit 
##                                          37 
##                                        Time 
##                                         156 
##                         Urban_or_Rural_Area 
##                                           0 
##                          Weather_Conditions 
##                                           0 
##                                        Year 
##                                           0 
##                                  InScotland 
##                                          53

Remove Column that Unrelated to This Research

traffic <- subset(traffic, select = -c(X1st_Road_Class, 
                                       X1st_Road_Number, 
                                       X2nd_Road_Class, 
                                       X2nd_Road_Number, 
                                       Location_Easting_OSGR, 
                                       Location_Northing_OSGR))

Remove Row that has NA value

traffic <- na.omit(traffic, cols = c("Did_Police_Officer_Attend_Scene_of_Accident", 
                                     "Longitude", "Pedestrian_Crossing.Human_Control", 
                                     "Pedestrian_Crossing.Physical_Facilities", 
                                     "Speed_limit", 
                                     "InScotland"))

Check The Table Again

traffic

4. Data Explanation

4.1. Brief Explanation

glimpse(traffic)

## Rows: 2,042,570
## Columns: 28
## $ Accident_Index                              <chr> "200501BS00001", "20050...
## $ Accident_Severity                           <fct> Serious, Slight, Slight...
## $ Carriageway_Hazards                         <fct> None, None, None, None,...
## $ Date                                        <date> 2005-01-04, 2005-01-05...
## $ Day_of_Week                                 <fct> Tuesday, Wednesday, Thu...
## $ Did_Police_Officer_Attend_Scene_of_Accident <fct> 1, 1, 1, 1, 1, 1, 1, 1,...
## $ Junction_Control                            <fct> Data missing or out of ...
## $ Junction_Detail                             <fct> Not at junction or with...
## $ Latitude                                    <dbl> 51.48910, 51.52007, 51....
## $ Light_Conditions                            <fct> Daylight, Darkness - li...
## $ Local_Authority_.District.                  <fct> Kensington and Chelsea,...
## $ Local_Authority_.Highway.                   <fct> Kensington and Chelsea,...
## $ Longitude                                   <dbl> -0.191170, -0.211708, -...
## $ LSOA_of_Accident_Location                   <fct> E01002849, E01002909, E...
## $ Number_of_Casualties                        <int> 1, 1, 1, 1, 1, 1, 1, 2,...
## $ Number_of_Vehicles                          <int> 1, 1, 2, 1, 1, 2, 2, 1,...
## $ Pedestrian_Crossing.Human_Control           <int> 0, 0, 0, 0, 0, 0, 0, 0,...
## $ Pedestrian_Crossing.Physical_Facilities     <int> 1, 5, 0, 0, 0, 0, 0, 0,...
## $ Police_Force                                <fct> Metropolitan Police, Me...
## $ Road_Surface_Conditions                     <fct> Wet or damp, Dry, Dry, ...
## $ Road_Type                                   <fct> Single carriageway, Dua...
## $ Special_Conditions_at_Site                  <fct> None, None, None, None,...
## $ Speed_limit                                 <fct> 30, 30, 30, 30, 30, 30,...
## $ Time                                        <Period> 17H 42M 0S, 17H 36M ...
## $ Urban_or_Rural_Area                         <fct> Urban, Urban, Urban, Ur...
## $ Weather_Conditions                          <fct> Raining no high winds, ...
## $ Year                                        <int> 2005, 2005, 2005, 2005,...
## $ InScotland                                  <fct> No, No, No, No, No, No,...

4.2 Summary

4.2.1. Highest Category in Each Variable

A: Accident Severity

traffic %>% 
  count(Accident_Severity, sort = TRUE)

B: Junction Control

traffic %>% 
  count(Junction_Control, sort = TRUE)

C: Junction Detail

traffic %>% 
  count(Junction_Detail, sort = TRUE) %>% 
  head(10)

D: Urban/Rural Area

traffic %>% 
  count(Urban_or_Rural_Area, sort = TRUE)

E: Road Type

traffic %>% 
  count(Road_Type, sort = TRUE)

F: Road Survey COndition

traffic %>% 
  count(Road_Surface_Conditions, sort = TRUE)

G: Speed Limit

traffic %>% 
  count(Speed_limit, sort = TRUE)

H: Weather Condition

traffic %>% 
  count(Weather_Conditions, sort = TRUE)

F: Light Condition

traffic %>% 
  count(Light_Conditions, sort = TRUE)

G: Year

traffic %>% 
  count(Year, sort = TRUE)

4.2.2. Highest number of Accident Based on Accident Severity

A: Year

traffic %>% 
  count(Accident_Severity, Year, sort = TRUE)

B: Day of Week

traffic %>% 
  count(Accident_Severity, Day_of_Week, sort = TRUE)

C: Local Authority

traffic %>% 
  count(Accident_Severity, Local_Authority_.District., sort = TRUE)

D: Light Condition

traffic %>% 
  count(Accident_Severity, Light_Conditions, sort = TRUE)

E: Speed Limit

traffic %>% 
  count(Accident_Severity, Speed_limit, sort = TRUE)

F: Road Type

traffic %>% 
  count(Accident_Severity, Road_Type, sort = TRUE)

G: Junction Control

traffic %>% 
  count(Accident_Severity, Junction_Control, sort = TRUE)

H: Junction Detail

traffic %>% 
  count(Accident_Severity, Junction_Detail, sort = TRUE)

5. Data Insight

CASE 1

TOP 5 THE CAUSE OF ACCIDENT

traffic %>% 
  filter(Accident_Severity == "Fatal") %>% 
  filter(Junction_Control != "Data missing or out of range") %>% 
  group_by(Urban_or_Rural_Area, Light_Conditions, Junction_Control, Junction_Detail) %>% 
  count(sort = TRUE) %>% 
  head(5)

Fatal Accident mostly happens in Urban area when sun still shines. T or Staggered Junctions with uncontrolled by the police are common place to have an accident

CASE 2

IS SCOTLAND MORE DANGEROUS?

traffic %>% 
  filter(Accident_Severity == "Fatal") %>%
  group_by(InScotland) %>% 
  count(sort = TRUE) %>% 
  head(5)

Scotland’s number of Fatal accident is almost a 1/10 than Outside them

CASE 3

TOP 5 OF ROAD TYPE AND ROAD SURFACE CONDITIONS WITH FATAL ACCIDENT

traffic %>% 
  filter(Accident_Severity == "Fatal") %>% 
  group_by(Accident_Severity, Road_Type, Road_Surface_Conditions) %>% 
  count(sort = TRUE) %>% 
  head(5)

Single Carriageway with Dry surface is mainly have a Fatal Accident

CASE 4

TOP 5 OF WEATHER CONDITIONS, LIGHT CONDITIONS AND SPEED LIMIT WITH FATAL ACCIDENT

traffic %>% 
  group_by(Accident_Severity = "Fatal", Weather_Conditions, Light_Conditions, Speed_limit) %>% 
  count(sort = TRUE) %>% 
  head(5)

A broad day with fine no high winds weather is most likely have Fatal Accident

CASE 5

TOP 5 DAYS OF FATAL ACCIDENT

traffic %>% 
  filter(Accident_Severity == "Fatal") %>% 
  group_by(Day_of_Week) %>% 
  count(sort = TRUE) %>% 
  head(5)

Fatal Accident most likely happens on Weekend

CASE 6

TOP 5 LOCAL AUTHORITY & POLICE FORCE THAT MANAGE FATAL ACCIDENT

traffic %>% 
  filter(Accident_Severity == "Fatal") %>% 
  group_by(Local_Authority_.District., Police_Force) %>% 
  count(sort = TRUE) %>% 
  head(5)

City of Birmingham under West Midlands Police Force handle the most accident from 2005 - 2017